A safe screening rule for sparse logistic regression

J. Wang, J. Zhou, J. Liu, P. Wonka, J. Ye
Advances in Neural Information Processing Systems 27, conference, pp. 1053-1061, (2014)

A safe screening rule for sparse logistic regression

Keywords

Sparse logistic regression

Abstract

​The`1-regularized logistic regression (or sparse logistic regression) is a widely used method for si- multaneous classi cation and feature selection.  Although many recent e orts have been devoted to its ecient implementation,  its application to high dimensional data still poses signi cant challenges.  In this paper, we present a fast and effective sparse logistic regressions creening rule (Slores) to identify the  \0"  components  in  the  solution  vector,  which  may  lead  to  a  substantial  reduction  in  the  number of features to be entered to the optimization.  An appealing feature of Slores is that the data set needs to be scanned only once to run the screening and its computational cost is negligible compared to that of solving the sparse logistic regression problem.  Moreover, Slores is independent of solvers for sparse logistic regression, thus Slores can be integrated with any existing solver to improve the eciency.  We have  evaluated  Slores  using  high-dimensional  data  sets  from  di erent  applications.   Extensive  experi- mental results demonstrate that Slores outperforms the existing state-of-the-art screening rules and the eciency of solving sparse logistic regression is improved by one magnitude in general.

Code

-

Sources

Website PDF

See all publications 2014